Shop OBEX P1 Docs P2 Docs Learn Events
spider bot? any biologists out there that no how there legs work? — Parallax Forums

spider bot? any biologists out there that no how there legs work?

mctriviamctrivia Posts: 3,772
edited 2010-04-12 23:54 in Robotics
Is there any reason other then cost why all the robot spider kits are hexapods(6 legged)?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out my brand new site.

Post Edited (mctrivia) : 4/10/2010 2:00:22 AM GMT

Comments

  • Martin_HMartin_H Posts: 4,051
    edited 2010-04-08 01:23
    Well six legs is the minimum number of legs needed to do the alternate tripod gate and be statically stable. I don't think adding two more leg buys you anything other than complexity and increased power consumption. Neither the robot's speed nor stability would be increased.

    I've seen some mechanical Lego designs that have eight legs powered by two motors, but I assumed that was a fallout of their gear trains and not a design goal.
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-08 03:06
    lgebr_spider.jpg

    I was just wondering because real spiders have 8 legs. Though I guess most kits only have 1 joint in the middle of the leg and it looks like mos of the real ones have 2. So maybe realism is just to expensive and complicated.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.
  • P!-RoP!-Ro Posts: 1,189
    edited 2010-04-08 03:55
    I think the real spiders used 8 instead of 6 to:
    1. allow them to climb on their web easier without breaking it
    2. increase attacking ability with 6 back legs doing the walking and the 2 front attacking.
    Also--nice to see you moved from web spider to robot! I think you should make it 8 legged just to see if you can mimic a spiders attack stance.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Pi aren't squared, pi are round. Cornbread are squared!
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-08 04:08
    well I am thinking of putting a propeller powered spider kit together. will need to see how cheap I can find servos.

    3 deg of freedom * 6 legs =18 servos
    3 deg of freedom * 8 legs =24 servos


    I do like the idea of having 8 legs like a real spider though.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.
  • P!-RoP!-Ro Posts: 1,189
    edited 2010-04-08 04:16
    If you use hobbypartz 9g micro servo it'll come out as $66.48. Not too bad, just sell a propmod and a half and you should have it (lol I know you don't get pure profit). It does seem like a lot of servos though, I'll admit to that.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Pi aren't squared, pi are round. Cornbread are squared!
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-08 04:23
    nope not pure profit. actually pretty low profit margine. But I am slowly expanding my inventory. Robots is where I am going next and I think spiders are cool.

    Those servos do not have the torque needed. Once I figure out the mass of the body parts I should be able to figure out how much torque is needed. hopefully under $10 servo will do.

    Thanks for the complement. I like my new site to. Looks cool and the built in inventory system makes my life a lot easier. Now if only I could get it to assemble, and ship the stuff for me also.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.
  • P!-RoP!-Ro Posts: 1,189
    edited 2010-04-08 05:07
    Strange you mention it here, I actually made the compliment in the other thread. Yeah, I can see your site advertising robotics and I think it would be quite cool. I would even love to buy one and try it out this summer--even if it was still a prototype. If nothing else the experience from programming a spider should help me in my ultimate goal--building a full out humanoid. Anyway, good luck and have fun! If robotics wasn't enjoyable I wouldn't spend over a year trying to get a single bot to run.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Pi aren't squared, pi are round. Cornbread are squared!
  • Robert TRobert T Posts: 71
    edited 2010-04-08 15:29
    I say do 8 legs, just because its more difficult, not typically done( I have never seen an 8 leg robot) and will look great if you pull it off.

    The front two legs could also be multipurpose rather than just for locomotion.
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-08 15:58
    I agree. Can always center wait over back 6 legs so front 2 are optional

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.
  • Spiral_72Spiral_72 Posts: 791
    edited 2010-04-08 16:42
    mctrivia said...
    Those servos do not have the torque needed. Once I figure out the mass of the body parts I should be able to figure out how much torque is needed. hopefully under $10 servo will do.


    Hobbycity.com has some very inexpensive servos. If I remember correctly, $4 buys a servo comparable to the HS-55. I used them in delta wing combat [noparse]:)[/noparse]

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "puff"...... Smile, there went another one.
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2010-04-09 13:51
    Go with 8 and put force feedback on the front 2. It could 'feel' it's way around.
    Cool idea!

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-09 15:03
    I realised that the kits out there are based on ants not spider. Spiders are cool though so let's do that.

    Current sensing on servos should be effective force feedback method right

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.
  • P!-RoP!-Ro Posts: 1,189
    edited 2010-04-09 21:50
    I'd use actual force sensors and bend them to a right angle to detect both grasping something and pushing.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Pi aren't squared, pi are round. Cornbread are squared!
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-09 22:18
    well i will have to look into what force sensors are available.

    i realised I can make both designs reasonably:

    ant:
    *6 legs, 3 servos/leg
    *head, 5 servos
    *body, 2 servos
    25 total

    spider:
    *8 legs, 4-6 servos/leg
    *head, 2 servos
    *body, 2 servos
    38-54 servos total

    since they both would use the same servos and controlers(though the spider would need 2 controlers)

    one thing I have been having some trouble with though is figuring out how the legs of the spider works:

    attachment.php?attachmentid=69302

    shows that there are 11 muscles in the spiders legs
    i. Extensor trochanteris, ii. Flexor trochanteris; iii. Flexor longus femoris; iv. Flexor bilobatus femoris; v. Flexor bilo- batus patellae ; vi. Flexor patellae ; vii. Protractor tibiae ; viii. Flexor tibiae : ix. Flexor metatarsi ; x. Extensor tarsi ; xi. Flexor tarsi.

    SPIDER1.gif

    shows that there are 8 parts to the leg:
    1. coxa
    2. trochanter
    3. femur
    4. patella
    5. tibia
    6. metatarsus
    7. tarsus
    8. tarsal claws

    I think the muscles work as follows:
    i,ii: provide left/right movement of the trochanter
    iii,iv provides up/down movement of the femur
    v,vi: provide up/down movement of the patella
    vii,viii: provide left/right movement of the tibia
    ix: up/down movement of the metarsus
    x,xi: up/down movement of the tarsus

    but this would mean i need 6 servos/leg. also i can't find any images showing the legs in any possition but straight(contradicting my vii,viii asumption) and the patella and tibia always seem to be straight like they are fused. that would make only 4 servos needed.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.

    Post Edited (mctrivia) : 4/9/2010 10:24:55 PM GMT
    370 x 435 - 19K
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2010-04-10 00:07
    For force sensors you could use current to the servo OR put some of those cool Boe-bot wisker sensors at the end of the leg. For feeling around for objects, you don't need to know how big/heavy/hard it is, just that it's 'there'. A binary sensor might be the trick.
    Jim-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
  • Robert TRobert T Posts: 71
    edited 2010-04-10 18:04
    McTrivia

    Treat i,ii,iii, iv as the hip joint: It needs up/down and forward/back
    Treat v, vi, vii, viii as the knee joint. You really only need flexion/extension since the knee does not abduct/adduct(right/left)
    Treat ix, x, xi as the ankle/toes. Similar to knee joint. Only flexion/extension needed. The ankle and foot have very little lateral movement
  • Robert TRobert T Posts: 71
    edited 2010-04-11 16:21
    Here is an interested video of an eight leg robot:



    My first impression was that the algorithm could be improved for walking. Second, eight legs makes for a lot more work to move the same distance as say a hexapod. Also with eight legs there is not much room to move laterally at the hip joint given the arrangement of the legs.

    Maybe for standard flat surface motility, use only six legs. The question is, under what situations are eight legs better than six. If the front two legs were longer than the others, they could be used for climbing, for example. They would be more like arms than legs.

    I could not find any video of slow motion spider locomotion.
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-11 17:11
    I found a slow motion video wll post in an hour.

    Real spiders can sacrifice legs to escape.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-11 18:18
    www.youtube.com/watch?v=dUneI-WOqVk&NR=1

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Check out my brand new site.
  • MacGeek117MacGeek117 Posts: 747
    edited 2010-04-12 17:53
    " I think you should make it 8 legged just to see if you can mimic a spiders attack stance."
    Check out the sample vids for the CH3-R hexapod from Lynxmotion.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    BS2: $49


    SX48 Protoboard: $10

    Propeller Protoboard: Priceless

    www.apple.com

    www.parallax.com

    www.goldmine-elec.com

    www.expresspcb.com

    www.jameco.com
  • Martin_HMartin_H Posts: 4,051
    edited 2010-04-12 23:38
    From that video it looks like spiders use an alternate tetra-pod gate.
  • mctriviamctrivia Posts: 3,772
    edited 2010-04-12 23:54
    Ya there gate is a lot nore complicated then an ants but it looks pretty cool.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    stock.gif
Sign In or Register to comment.